Search research

Dynamic black-box optimization presents significant challenges for Bayesian Optimization (BO), as the objective function evolves over time, causing optimal locations to shift continuously. Existing dynamic BO (DBO) methods using standard acquisition functions such as Upper Confidence Bound (UCB) fail to explicitly account for temporal variations, leading to suboptimal sample allocation and poor tracking of moving optima. Here, we propose BOBA (Bayesian Optimization through Bayesian Active Inference), a novel acquisition function inspired by free energy principles from active inference that explicitly minimizes predictive uncertainty about future states in dynamic environments. BOBA extends traditional acquisition functions by incorporating a forward-looking uncertainty quantification that estimates uncertainty in function changes, enabling more informed exploration-exploitation trade-offs in non-stationary settings. We evaluate BOBA on synthetic dynamic benchmarks, comparing against state-of-the-art DBO methods. Our experiments demonstrate that BOBA significantly improves regret in query-restricted settings, while remaining competitive in time-limited settings. We further analyze variants of BOBA with different exploration strategies, showing how the exploration-exploitation balance can be tuned for different types of dynamic functions. This work contributes both a free energy-based acquisition function for DBO and insights into how active inference principles can enhance optimization in non-stationary environments, with implications for real-time applications requiring continuous adaptation.

Mixed-precision weight quantization is commonly formulated as a Multiple-Choice Knapsack Problem (MCKP), yet existing solvers rely on scalar sensitivity proxies that collapse each weight matrix's Hessian into a single number and treat every module independently. We prove that even the optimal scalar proxy incurs multiplicative distortion up to $\sqrt{κ(\mathbf{A})κ(\mathbf{B})}$ relative to the full activation-aware quadratic, where $κ(\mathbf{A})$ and $κ(\mathbf{B})$ denote the condition numbers of the input- and output-side Hessian factors. This bound varies from $10^1$ to $10^{13}$ for typical LLM modules, making inter-module sensitivity ranking unreliable. To address these limitations, we propose Cross-layer Activation-aware Sensitivity Allocation (CASA), a two-phase method. In Stage 1, the scalar proxy is replaced by an activation-aware metric derived from the Kronecker-factored Hessian, reducing the MCKP to a form whose continuous relaxation admits a closed-form solution. In Stage 2, a cross-layer-aware local search evaluates bit-width updates using the end-to-end model loss. Experiments on multiple LLMs across different bit budgets show that CASA achieves lower perplexity than the latest scalar-proxy baselines, especially at ultra-low bit-widths ($<3$ bits per weight). Moreover, the performance gain in zero-shot accuracy tracks the per-model average condition-number over modules, confirming the distortion bound as a practical indicator of scalar-proxy failure.

When we evaluate the performance of our odometry, it is common practice to score the estimated track against a ground truth. Unfortunately, scoring uses point metrics, such as the root mean square error, that ignore the covariance matrix which estimators like filters and smoothers already report. Using the covariance matters for two reasons. First, the covariance encodes the estimator's uncertainty, so it tells us whether the estimator trusts its own output. An overconfident estimator will not report itself lost. Second, the covariance weights the error in each direction of the estimate. Without the covariance, an estimator is unduly penalized for a high error in an uncertain direction. Instead of point metrics, we should use strictly proper scoring rules. These rules score the estimate together with its reported uncertainty. Strictly proper scoring rules recover the point metrics when no covariance is reported, and they diagnose covariance inconsistency when covariance is reported. Using a one-sided pairwise test, we show that two estimators can expose overconfidence in at least one of them without a ground truth. Strictly proper scoring rules and our pairwise test are available in our open-source framework smfeval. As a case study, we use smfeval to assess the uncertainty quality of the translational component of ground-based LiDAR-inertial odometry. Across four filters we find overconfidence - the worst case reports centimeter certainty with kilometer error. Knowing the filters are overconfident, we investigate the mechanism. The investigation traces overconfidence to filters crediting LiDAR measurements with more new information than they carry.

Aerial-ground cooperation requires real-time UAV--UGV relative-state information. Instead of maintaining global estimates for both robots, direct control in a UGV-attached non-inertial frame avoids reliance on global localization. Vision-based relative pose estimation with a passive marker offers a low-cost and effective solution. However, a fixed camera may lose sight of the moving UGV when the required UAV attitude conflicts with the field-of-view (FOV) constraint. To address this, we propose COPA, a robust active-perception framework for global-state-free aerial-ground cooperation. We use a single-axis gimbal to decouple the camera optical axis from the UAV pitch attitude. We derive an active-perception model that relates UAV motion, gimbal angle, and UGV motion to the target image-plane state.A Temporal Convolutional Network (TCN) predicts short-horizon UGV acceleration and angular velocity from recent motion history without global-state measurements. The model predictive control (MPC) uses these predictions to jointly optimize UAV and gimbal control. Simulations show that COPA maintains continuous target visibility, while ablation studies confirm that the TCN reduces peak errors during UGV motion transitions. Real-world experiments with UGV accelerations up to 3m/s^2 and yaw rates up to 1.0rad/s demonstrate robust tracking.

Manipulation requires both broad scene awareness and detailed local feedback, yet conventional camera rigs provide them through separate front and wrist cameras. We present Fisheye-VLA, a visual interface that brings these capabilities together using a single passive fisheye. A global view preserves the workspace, while local perspective crops direct detail toward the interaction. The key design question is where this local visual budget should go. We answer it through a controlled re-rendering study, comparing alternative crop directions on the same recorded observations. The study finds that end-effector-centered views capture most of the estimated benefit of a much larger candidate pool, motivating a compact allocation around both hands. Our interface uses calibrated end-effector projection and motion lead to track the crops, while a shared ray encoding preserves their spatial meaning as they move. Integrated with a pretrained VLA, it achieves 84% and 82% success in the two expanded tabletop regions, where some target placements extend beyond the front-camera coverage, and supports shelf and conveyor manipulation. Ablations show that local crops and their viewing directions become more important in the larger workspace regions. The results demonstrate that a single fisheye can support these manipulation tasks without physical wrist cameras.

An animal with a weakened limb does not necessarily switch its gait, instead it unloads the affected limb, re-coordinates the remaining limbs, and scales its response with injury severity. This graded adaptation allows locomotion to persist despite partial loss of limb strength, rather than requiring a discrete transition between healthy and failed. Inspired by this behavior, we propose SG-CPG, a central pattern generator (CPG) for quadruped locomotion under continuous actuator degradation. SG-CPG preserves a frozen healthy CPG policy and introduces two severity-driven gates: a residual gate that re-coordinates all four legs and an amplitude gate that progressively shortens the weakened leg's stride as degradation increases. We emulate progressive degradation through two mechanisms: lowering the joint torque ceiling (ceiling mechanism) and scaling its low-level controller gains (gain mechanism), representing distinct forms of actuator weakening. Our simulations on a Unitree Go2 show that SG-CPG maintains a trot gait with 100% survival across an omnidirectional command schedule under 95% joint strength loss while tracking commands within 8%. Under a lowered torque ceiling, removing either severity path, the residual's severity observation or the amplitude gate, raises clipping at the weakened joint from 4.4% to 13.6% and 26.3% of steps at an 80% loss. On a real Go2, SG-CPG survives 28 of 29 forward and turning trials with up to 93% calf torque degradation. These results show that severity-gated adaptation can extend a healthy locomotion policy to progressive actuator degradation without treating the fault as a discrete failure.

Long-horizon assigned work requires an LLM agent to track the state of a task: which steps are done, blocked, cancelled, or open to repetition. Agent systems either keep this state as text in the prompt and rely on the model to read that text, or move the state into a module that enforces it, and each system is evaluated as a whole, so no one knows how much reliability comes from the state being shown, told, or enforced. We fix the task rules, the model, and paired episodes and vary how strongly task state reaches the agent: a raw transcript, an exact checklist, per-turn directives from a state machine compiled from the brief and advanced only by execution receipts, or an enforcement gate on that machine that refuses state-violating actions; every episode is scored by exact payload matching against dynamic ground truth. Across three models, two reasoning regimes, and two domains, four findings hold without per-turn reasoning: displaying accurate state is unreliable, an unverified ledger the agent writes itself beats an accurate checklist it is shown, directives help in proportion to the model's obedience, and enforcement needs no obedience but is bounded by the correctness of its state and by the matcher that maps requests to steps; per-turn reasoning at a 235B agent compresses these separations without repairing the text rungs. The same gate, compiled from $τ^2$-bench's airline policy, raises a 235B agent's pass$^1$ from 0.39 to 0.54 and changes nothing for a 35B agent that rarely violates the policy; on PM-Bench, where acting turns on recognizing a cue rather than on state, showing the record is the best rung--matching or beating both gates and reversing the ledger-over-checklist finding--and enforcing the matcher's judgement drops a 35B agent below its raw transcript. Enforcement pays when failures are state-decidable and frequent, and hurts when the gate's judgement is wrong.

Lower cost open source robots and reinforcement learning (RL) simulation tools create new opportunities for precollege students to engage with contemporary robotics. However, translating a complete research workflow, spanning mechanical assembly, electrical setup, simulation, policy learning, system identification, and physical deployment, into a coherent course for novice learners remains challenging. We present an integrated robotics course framework that organizes these activities around a shared robotic artifact. The framework combines parallel disciplinary tracks, sequencing based on technical dependencies, progressive integration of simulation and hardware, layered performance checkpoints, and structures for balancing collaborative work with individual accountability. We illustrate the framework through a high school curriculum organized around a robot project in which pairs of students assemble an open source humanoid robot, train a walking policy in simulation, and deploy it on the physical platform. The framework was developed through an iterative design process that included formative review by five experts in robotics research, engineering, secondary STEM education, and curriculum design. Expert feedback highlighted three central design tensions: authenticity versus cognitive load, system integration versus timely visible progress, and team construction versus individual accountability. These tensions informed the final framework presented in this paper. This work offers a structured approach for adapting robotics research workflows into interdisciplinary precollege courses; future classroom studies are needed to examine implementation and student learning.

Real-world reinforcement learning (RL) systems still struggle with the demands of contact-rich industrial manipulation, including micrometer-level precision, success rates above 99%, and human-level cycle times. Although off-policy algorithms can improve performance by leveraging demonstrations and interventions, a key bottleneck is the lack of an intuitive interface for collecting such guidance while complying with constraints of the physical system and the policy. We propose PAKT, a framework for kinesthetic teaching in RL. As opposed to teleoperation approaches, PAKT relies on kinesthetic guidance, which is widely used in industry. However, a critical weakness of kinesthetic guidance is the possibility for the operator to move the robot along trajectories (e.g., velocities, accelerations, jerk) that the robot and/or policy cannot physically reproduce. Using PAKT, operators guide the robot through admittance control, which maps human-applied forces to motion. The downstream reference generator applies the same kinematic limits used during policy execution, keeping the collected trajectories within these limits. To support this teaching interface with an appropriate execution layer, PAKT adds a high-performance control stack that maps low-frequency RL actions to high-frequency torque commands. It consists of a reference generator and subsequent impedance controller, where the reference generator preserves the tracking performance of the impedance controller while improving contact handling and producing smoother policy actions. Across the reported runs on four insertion and industrial assembly benchmarks, including a data center compute tray, the end-to-end system reduces cycle time by 23%-48% and cumulative intervention count by 62%-86% relative to the HIL-SERL baseline. Project website: https://pakt-website.github.io/pakt-website}{https://pakt-website.github.io/pakt-website

Most minimally invasive procedures are still performed with hand-held laparoscopic instruments, yet only the endoscopic video is retained; the instrument motion that expresses surgical skill, and that could support skill assessment and robot learning, is lost. Pose from video alone remains millimeters to centimeters off, and an instrument-mounted inertial measurement unit (IMU) cannot rely on its magnetometer, whose field changed with tool pose and between sessions in our measurements. We present a surgical instrument-state logger that clips onto a conventional instrument without modifying the part that enters the patient and fuses a six-axis IMU and a time-of-flight (ToF) rangefinder with a markerless camera in an error-state Kalman filter under the remote center of motion (RCM) of the trocar. Heading comes from the shaft silhouette, segmented by a U-Net, in place of the magnetometer: the rotation-angle error is 0.200°, against 3.58° from the accelerometer and magnetometer alone. Against a Franka Research 3 manipulator, and without alignment to it, the displacement error over 300 translation trials was 1.21mm RMS and the relative-rotation error over 180 rotation trials 0.34° RMS. On continuous trajectories, tracked and displayed in real time, the absolute tip error was 1.22mm (programmed) and 3.04mm (teleoperated) after post-hoc tuning of three filter parameters, and the full fusion beat every sensor subset. Because the estimator uses no magnetic measurement, its accuracy does not rely on an undisturbed field. The same clip-on device could thus record metric tip trajectories during routine hand-held laparoscopy, while displaying the insertion depth and attitude that are hidden once the instrument is inside the patient.

Soccer tactics are interactive: an attacking action changes the opponent's defensive problem, and the observed response depends on the multi-agent match state. We introduce SambaGraph, an action--reaction spatio-temporal graph dataset and benchmark for soccer tactical response modeling. From tracking and event data for all 64 matches of the 2022 FIFA World Cup, we curate 4,070 action-centered episodes represented as temporally aligned 23-node player--ball graph sequences with attack/defense views, response labels, and 26,270 split-safe attack--defense pairs. We study three questions: whether observed responses can be classified from graph episodes, whether successful defenses can be retrieved for a query attack, and whether graph-derived summaries support grounded LLM reasoning. A compact signature MLP obtains $0.796\pm0.007$ macro-F1 for response classification, while a fused graph--signature dual encoder reaches $0.471\pm0.029$ Hit@5 and $0.655\pm0.051$ Hit@10 for full-bank defensive retrieval. Hard negatives maximize pair discrimination but not retrieval quality. Local LLMs underperform supervised encoders for direct classification and do not improve over a strong original order in eight-candidate reranking, but they provide grounded tactical rationales. These results position SambaGraph as a reproducible benchmark for graph-based soccer strategy-response research. Code and dataset are available at: https://github.com/areyesan/SambaGraph.

Current robot-assisted minimally-invasive surgery (RMIS) platforms provide a fixed console for the surgeon to view stereo endoscopic images and teleoperate instruments inside the patient. Several researchers have proposed the use of a head-mounted display (HMD) as a portable console, with video pass-through rendering of the endoscope images which, like the fixed console, restricts the operator to a single endoscopic viewpoint and limits depth perception. We present a digital twin-driven virtual reality (VR) teleoperation platform, where the digital twin is created from markerless perception of the surgical environment and streamed for display on the HMD. This overcomes the limitations of video pass-through by providing multi-view rendering and natural motion-parallax cues, enabling decoupling of the user's hand posture from strict instrument alignment. The system utilizes VR hand controllers to increase the teleoperation workspace and to improve the robustness and stability of instrument control compared to the hand tracking approach adopted by most prior systems. A 15-participant user study on the da Vinci Research Kit (dVRK) shows that our VR platform significantly outperforms a state-of-the-art HoloLens 2 mixed reality baseline, reducing path length by 86% and jerk by 95%, while achieving depth perception confidence comparable to or exceeding the traditional console across all conditions.

Rapid intensification (RI) remains one of the most consequential and difficult aspects of tropical cyclone (TC) forecasting. Although full-physics numerical weather prediction models can represent the processes governing RI, resolving storm-environment interactions remains computationally expensive, while purely data-driven approaches often lack physical interpretability. We present FAST-ML, a hybrid framework that bridges data-driven efficiency with physical constraints. A physically informed dual-stream neural parameterization ingests 3D ERA5 fields to diagnose ventilation controls---environmental wind shear and mid-level entropy deficit. By optimizing these parameters end-to-end through a differentiable FAST intensity model, this architecture establishes a robust new paradigm for observation-driven parameter optimization, ensuring storm evolution remains strictly governed by thermodynamic principles. By better capturing the storm's continuous intensity evolution, FAST-ML improves upon its physical baseline, reducing ensemble CRPS across forecast lead times, with a reduction of approximately 31% at 60 h and nearly halving the RI false alarm ratio without sacrificing detection skill. In a 100-member ensemble configuration, FAST-ML produces intensity forecasts comparable to FNV3 for selected storms under the evaluated input configurations. Furthermore, zero-shot tests on selected Eastern Pacific storms provide encouraging evidence of cross-basin transferability. FAST-ML provides a modular intensity forecasting framework that can be coupled with externally supplied storm tracks and environmental fields. It demonstrates that observation-driven parameter learning within physically constrained dynamics simultaneously enhances accuracy, interpretability, and computational efficiency.

Consistent multi-view object segmentation is critical for 3D perception and robotics, yet remains challenging under severe viewpoint and occlusion changes. Existing methods typically perform 3D instance segmentation on point clouds or rely on offline 2D mask-matching pipelines. However, 3D instance segmentation is limited by scarce 3D annotations, while offline 2D matching suffers from object identity ambiguity across frames. To leverage strong 2D and 3D priors jointly, we propose SAM-V (Geometry-Aware Segment Anything for Multi-View Instance Segmentation). Instead of combining the two priors through post-hoc matching, SAM-V directly integrates features from a feed-forward geometry model (VGGT) into a 2D segmentation foundation model (SAM), trained end-to-end for cross-view instance prediction. SAM-V introduces a prompt-fusion mechanism that enriches sparse SAM prompt tokens with view-specific camera tokens and local VGGT features, making the prompt representation both view-aware and spatially grounded, together with a mask decoder that attends to dense 2D and 3D features. By conditioning the mask decoding directly on multi-view geometry, SAM-V produces consistent multi-view segmentation of a prompted object in a single forward pass without offline mask matching or explicit 3D reconstruction. On the IGGT 3D tracking benchmark, where consistent instance identity across frames directly determines performance, SAM-V improves overall IoU by 5 points and frame-level recall by 12 points on the ScanNet++ split over the state-of-the-art multi-view instance segmentation baseline and leads on all metrics in the zero-shot ScanNet split. Our code and pretrained models are available at https://github.com/gong208/SAM-V.git.

Forceful manipulation is challenging for humanoid robots because interaction forces can disturb whole-body balance. We introduce the Supporting Hand Strategy (SHS), which enables a humanoid to brace against the environment with one hand while performing forceful manipulation with the other. SHS optimises a task-conditioned support configuration that guides two synchronous reinforcement-learning policies, without human motion data or online whole-body trajectory planning. On a Unitree G1, SHS achieved usable contact forces up to 60 N, compared with a maximum sustained force of 13.5 N without environmental bracing, while substantially improving force tracking over a task-independent support configuration. The same policies generalised to different task regions without retraining. SHS therefore provides a simple mechanism for substantially extending humanoid forceful-manipulation capability.

Bioinspired feedback controls for pursuit, tracking, and collective motion are often expressed in terms of the relative configuration between interacting agents. In practice, however, onboard sensors may not directly provide all quantities required for feedback control, necessitating estimation of unobserved quantities. This paper develops a bioinspired internal model-based estimator for reconstructing those quantities from partial sensory observations and known self-motion. State reconstruction is posed as an optimization problem that treats the relative kinematics as constraints and minimizes the disagreement between the internal model outputs and measurements from onboard sensors. Pontryagin's Maximum Principle is used to derive the necessary optimality conditions. A forward-backward algorithm is used to provide a numerical solution and a moving horizon formulation is employed for online implementation. The estimator is evaluated numerically against classical state estimators. Real-time implementation of the proposed framework on robotic hardware is demonstrated through two pursuit strategies.

Hand gesture recognition is a key component in human-computer interaction (HCI), enabling intuitive interfaces for applications in gaming, virtual reality (VR), robotics, and more. This study integrates transformer-based machine-learning models for real-time hand gesture recognition, using hand-tracking data captured through the OpenXR standard in Unity. We leverage positional data of hand joints and wrist rotation angles to train a custom gesture recognition system. By utilizing the sequential modeling capabilities of transformers, the system captures temporal dependencies within short gesture windows and classifies gestures robustly across hand orientations and sizes. The results show a significant improvement in gesture classification accuracy. Building on this, we outline how the approach can be extended toward detecting the flow of movement, i.e., the transitions between gestures, as future work.

Chain-of-thought (CoT) monitoring is only meaningful if written reasoning causally constrains the answer. We introduce continuation-based causal testing, an ablation-patch intervention that perturbs one reasoning step, truncates the chain, and forces the model to continue from the corrupted prefix. It measures how load-bearing a CoT is for the final answer, a behavioral notion distinct from mechanistic faithfulness. Across Gemma-2-9B-IT, Llama-3.1-8B-Instruct, and DeepSeek-R1-Distill-Qwen-7B on GSM8K, MMLU, and BIG-Bench Hard, CoT load-bearingness tracks model-relative task difficulty: on easy tasks models silently bypass their own reasoning; on hard tasks they follow corrupted steps and propagate errors. A matched 2x2 analysis shows task difficulty dominates perturbation type: error propagation rises 16x from GSM8K to BBH multistep arithmetic, and a variance partition over 28,584 continuations attributes 98.8% of explained deviance to task difficulty versus 0.8% to perturbation type. Reasoning-specific RL suppresses error propagation and compresses the gradient. A four-variant judge-sensitivity analysis and blind two-annotator study (n=500) show the error-propagation vs. non-propagation label is invariant to judge prompt, with perfect inter-annotator agreement (Cohen's kappa = 1.00). This gradient creates a structural problem for CoT-based oversight and AI safety monitoring: where the trace is easy to read it carries little signal, and where it matters errors propagate before a monitor can intervene. Linear probes on hidden states separate silent bypass, self-correction, and error propagation, but additive activation steering provides limited causal control, flipping only about 25% of error-propagation cases at best. Behavioral mode is readable but not reliably controllable.

Neural inertial odometry increasingly uses networks as learned measurements inside filtering pipelines. Such measurements should transform consistently under arbitrary IMU mounting conventions: their mean must transform as a vector, and their covariance must transform congruently as a second-order tensor. We present GINIO, a geometric SO(3)-equivariant interface for neural inertial odometry under arbitrary rotations of the IMU measurement frame. Given calibrated IMU windows, our framework predicts a motion measurement and uncertainty obeying these tensorial laws. To support efficient sensor-frame learning, we introduce Last-Frame Alignment (LFA), a deterministic preprocessing step that is provably equivalent to world-frame training for SO(3)-equivariant predictors. The connected estimator tracks sensor-local states such as IMU bias, separating nuisance estimation from the geometric law enforced by the learned measurement. We instantiate the same interface in filter-connected NIO, AirIO-style recurrent aerial prediction, EqNIO-style full-SO(3) canonicalization, and ResNet-style temporal backbones. On TLIO, GINIO achieves 2.018 m ID/SO(3) ATE while EqNIO degrades to 76.389 m, using 11.6x fewer FLOPs. On NanoBench, our AirIO-style instantiation improves ATE from 5.579 m to 1.430 m without external attitude input, and our ResNet-style instantiation reaches 0.581 m ATE versus 0.645 m for ResNet1D. On Fetch, GINIO empirically reduces unseen physical-remount ATE from 8.15 m to 0.50 m without retraining, demonstrating robustness beyond the exact coordinate-frame guarantee. For uncertainty, spectral covariance reduces covariance-equivariance error by over three orders of magnitude compared with a diagonal head.

Dexterous grasping requires deciding where to grasp, reaching the target, and maintaining stable contact. We connect these stages through a compact three-point interface that separates global geometric reasoning from local contact control. Given object geometry and optional language commands, our framework samples contact triples from a precomputed grasp-affordance heatmap. A model-based reactive controller tracks the object, avoids collisions, and guides the hand toward the selected contacts. In the final centimeters, a Reinforcement Learning (RL) policy uses proprioceptive feedback to refine and stabilize the grasp despite reaching and perception errors. It observes only finger joint states and its recent actions, with no target points, visual observations, or object geometry, so a single policy is shared across objects and grasp configurations. In simulation, we compare grasp-and-lift success against squeeze and end-to-end baselines, characterize reaching convergence, and demonstrate grasp steering; hardware demonstrations on two training objects and one unseen object illustrate the full pipeline. Our modular framework uses geometry to guide the reach and local feedback to secure the grasp.

Continual learning with Low-Rank Adapters (LoRA) typically mitigates forgetting by penalizing the overlap between a new update and the accumulated past weights, which discourages certain update directions without controlling how an update distributes its energy over the ones that remain. We ask whether that restriction has to be task-aware, or whether a generic one supplied by the optimizer is enough. We train a plain incremental LoRA (IncLoRA) with Muon, which orthogonalizes each update, and compare it against O-LoRA and ELLA over five seeds and three task orders on the Standard CL Benchmark and three seeds on TRACE. IncLoRA+Muon reaches the accuracy band of the dedicated methods on Standard CL and improves on every AdamW configuration on TRACE. One update-constraining mechanism is enough, whether it comes from the loss or from the optimizer; on Standard CL a second one does not help, and for the most restrictive method it costs 8.4 points of accuracy and the plasticity to fit each task. What separates the two optimizers is not the size of the update, which under Muon is 0.91 to 2.06 times that under AdamW, but how it is distributed. AdamW confines it to between 1.4 and 1.8 effective singular directions, Muon spreads it over 7.0, and the two do not overlap in any tracked run. Part of the advantage usually attributed to dedicated CL methods may therefore be explained by the geometry of the optimizer's updates.

Human demonstrations offer a scalable way to collect manipulation data, but their contacts may be unstable or infeasible when transferred to a robot hand. Collecting demonstrations directly on the target robot avoids this mismatch but substantially increases the cost of data collection. To address this trade-off, we present Touch2Robot, a framework that lets humans collect demonstrations while seeing how the target robot hand would contact the object. We capture human hand motion, tactile-glove measurements, and object motion during human manipulation. These recordings guide object-specific RL policies to reproduce the demonstrated object motion while favoring contacts consistent with the recorded human touch. We distill the learned behaviors into a unified real-time retargeter that maps incoming human observations and object geometry to robot hand configurations. During collection, the predicted robot configuration is synchronized with the tracked object pose in simulation to reconstruct robot-object contacts, which are visualized to help the demonstrator adapt subsequent interactions to the target hand. Across four real-world tasks, Touch2Robot improves average real-robot replay completion from 37.9% to 72.1% over visual-only feedback, while reducing the collection time per replay-successful demonstration from 58.6s to 18.2s. Reconstructed target-hand contacts achieve 44.2% F1 against real-robot tactile measurements, and policies trained on Touch2Robot demonstrations improve downstream Diffusion Policy performance by 29.1 percentage points over visual-only feedback. These results show that bringing robot touch into the human demonstration loop improves both the quality and efficiency of scalable dexterous data collection. Project webpage: https://Touch2Robot.github.io/.

Modern Vision-Language Navigation (VLN) models rely mostly on pre-trained large Vision-Language Models (VLMs) to predict navigation actions. While this fusion of language instructions and visual observations allows multimodal reasoning, it obscures how information is routed across modalities or what mechanisms drive navigation decisions. Thus, it remains unclear whether VLN models ground their predictions in relevant semantic cues or can track task progress. In this work, we study the interpretability and steerability of VLN models. We use intervention-based metrics that measure how visual observations, instructions, and visual memory causally influence navigation decisions. Our results show that these navigation policies are sensitive to all input modalities and do not depend on a single one. We further show that these agents encode navigation progress and retain semantic structure from their VLM backbones, enabling concept-level steering through internal activations. Finally, we extract activation vectors for abstract behaviors to transfer them zero-shot to out-of-distribution real-world scenarios, improving performance without additional fine-tuning.

Diffusion data-point unlearning is typically evaluated immediately after each deletion, even though subsequent requests may repeatedly update the same model. We identify sequential reappearance, a failure mode in which an instance that is initially judged to be forgotten later returns to the memorized regime without reuse of the deleted data or adversarial fine-tuning. To capture this behavior, we introduce a target-level evaluation protocol that tracks whether each target is forgotten immediately, remains forgotten at the end of the sequence, or reappears during subsequent deletions. We further find that targets that later reappear exhibit sharper local denoising-loss geometry after deletion than targets that remain forgotten.

The paper proposes a novel data-driven framework for designing and training a feedback linearizing controller by explicitly incorporating relative degree based conditions into the learning process. This enables the conventional feedback controller components to be replaced by neural Lie derivatives, thereby facilitating a fully data-driven feedback linearization framework. Furthermore, practical closed-loop stability is established by deriving sufficient conditions under which bounded identification errors lead to bounded tracking errors. The derived theoretical results are validated through their application to an armature controlled DC motor.

Climate variability influences whether a market disruption escalates into a food crisis, yet broad climate patterns like El Niño, tracked months before they alter hydro-climatic conditions, are still not incorporated as an early-warning component in food-security responses. We address this gap by introducing sensitivity regimes, a stratification of regions by the direction and strength of their vegetation response to the El Niño Southern Oscillation, and using them to estimate how food price spikes affect acute food insecurity across sub-Saharan Africa. Integrating remote sensing, socioeconomic data, and causal machine learning, we find that in regions where ENSO systematically suppresses vegetation, a price spike raises the share of the population at acute risk by 5.4 percentage points in the following month. In regions where vegetation is unaffected by or positively linked to ENSO, the estimated effect is smaller (around 2 percentage points) and statistically insignificant. These results demonstrate that climate context is critical for understanding food security vulnerabilities. Sensitivity regimes can be combined with operational price-spike triggers to stage anticipatory action: the ENSO state flags vulnerable regions months ahead, and a pre-positioned response in those regions to a price spike would avert the largest jump in acute food insecurity.

We ask how little geometry a person has to draw to control both where people stand and where the camera moves in a generated video. Our answer is a ground plane and one cylinder per person. A user draws a grid on the ground, places one cylinder where each person should stand, moves the cylinders and the camera over 81 frames, and the model renders a photoreal video in which the people occupy the cylinders' positions, move as the cylinders move, and are seen from the drawn camera. Appearance comes from a text prompt and a background reference image; layout and motion come from the geometry. Because no dataset pairs such a signal with video, we build the pairs ourselves: an automatic engine writes 2000 captions from a combinatorial seed, generates a clip for each with a text-to-video model, and lifts every clip back to its geometry with person tracking, background inpainting, an agentic ground-mask loop, feed-forward multi-view reconstruction and a plane fit, with no real footage and no manual labels. A LoRA on Wan2.2-Fun-Control trained on 1935 such tuples follows drawn layouts and camera paths on hold-out clips: the generated people match the cylinders' count, order, position and height, the text changes who they are, the reference image changes where they are, and dolly-in, orbit, pan and crane paths are followed, dolly-out only weakly.

Evaluation scores used around LLM systems -- including reward models, rerankers, and LLM judges -- can track surface form instead of the quality they claim to measure. When presented with a terse correct solution and a commented buggy solution for the same MBPP problem, a public preference reward model selects the correct one no better than a coin flip (0.507). Subtracting the predictable surface component from such scores is increasingly common, but removal alone does not yield a more valid measurement: the removed component may carry construct-relevant signal, and residualization cannot tell which is which. Under designed interventions -- unit-test labels with comment-only edits -- residualization attenuates the reward model's format effects by about 0.12 on both correct and buggy code, while the correct-versus-buggy margins move by less than 0.01. In observational NLI and QA settings, we freeze a held-out replication before scoring and re-evaluate it using labels from disjoint annotators; this supports only a narrower conclusion: better agreement with the construct labels on a pre-declared slice where a surface-only predictor errs, not a repaired score. Full-population agreement falls in every observational setting with a reported positive slice gain, and within-question ranking falls in every such QA setting. When construct and surface features are entangled, residualization can decorrelate a score while degrading construct alignment, and, in a controlled model, configurations just as damaging to construct alignment pass every pre-adjustment check, so no committed gate is a guarantee. We assemble these distinctions into a reporting protocol whose outcomes, refusal included, state what an adjusted score may be claimed to show: an audit-time diagnostic reported beside the construct-alignment cost it incurs, never a replacement for the raw score.

In many real-world planning tasks, agents must tackle uncertainty about the environment's state and variability in the outcomes of any chosen policy. We address both forms of uncertainty as a first step toward safer algorithms in partially observable settings. Specifically, we extend Distributional Reinforcement Learning (DistRL)-which models the entire return distribution for fully observable domains-to Partially Observable Markov Decision Processes (POMDPs), allowing an agent to learn the distribution of returns for each conditional plan. Concretely, we introduce new distributional Bellman operators for partial observability and prove their convergence under the supremum p-Wasserstein metric. We also propose a finite representation of these return distributions via psi-vectors, generalizing the classical alpha-vectors in POMDP solvers. Building on this, we develop Distributional Point-Based Value Iteration (DPBVI), which integrates psi-vectors into a standard point-based backup procedure-bridging DistRL and POMDP planning. By tracking return distributions, DPBVI lays the foundation for future risk-sensitive control in domains where rare, high-impact events must be carefully managed. We provide source code to foster further research in robust decision-making under partial observability.

Intermittent visual loss disrupts target-relative feedback during underwater orbiting, making it difficult to maintain coordinated motion and reacquire a moving target. We present AquaOrbit, a reinforcement-learning controller with a recovery module for underwater target orbiting under interrupted visual feedback. During detection loss, the recovery module uses latched line-of-sight, roll, and depth references to support stabilization and target reacquisition. We train the controller in Isaac Sim with dynamics, observation, and vision-loss randomization. Evaluated without retraining in Gazebo/ROS2 under a different physics engine and perception perturbations, AquaOrbit completes 20/20 orbiting trials in each of the static- and moving-target conditions on an unseen variable-depth 3-D trajectory. In the moving-target condition, it reduces mean line-of-sight error by approximately 46% relative to a PID-based visual servoing controller with recovery while maintaining comparable path-tracking accuracy; removing the recovery module reduces completion to 9/20. Zero-shot physical deployment with fully onboard perception and control demonstrates elliptical, figure-eight, and variable-depth circular trajectories, including the latter two trajectory types absent from training. The robot maintains attitude stability during manual occlusions lasting up to 8s and reacquires the target within 2.5s in the reported attitude-induced field-of-view loss events.